Analysis and code archive: Understanding Modality-Wise Completeness Estimation in Language-Dominated Multimodal Sentiment Analysis under Random Input Missingness

Machine-readable analysis and code archive for a PeerJ Computer Science submission. The archive contains raw evaluation CSVs for CMU-MOSI and CMU-MOSEI, per-test paired-comparison statistics with Holm corrections, completeness-estimator calibration audits, evaluation-mask sensitivity analyses, efficiency benchmark outputs, sensitivity power analysis, figure files with provenance, checkpoint manifests, code snapshots of the modified LNLN implementation including the completeness-free control, analysis scripts, training/evaluation logs, configuration files, and dataset/checkpoint manifests. Checkpoint binaries are not included because of their size; their SHA-256 hashes, sizes, best epochs, and validation MAE are recorded so that retrained checkpoints can be verified. Dataset feature files are not redistributed; checksums of the exact feature files are recorded.